Use immediate_commit to benefit the primary OSD

osd.cpp

@@ -68,6 +68,10 @@ osd_t::osd_t(blockstore_config_t & config, blockstore_t *bs, ring_loop_t *ringlo
     osd_num = strtoull(config["osd_num"].c_str(), NULL, 10);
     if (!osd_num)
         throw std::runtime_error("osd_num is required in the configuration");
+    if (config["immediate_commit"] == "all")
+        immediate_commit = IMMEDIATE_ALL;
+    else if (config["immediate_commit"] == "small")
+        immediate_commit = IMMEDIATE_SMALL;
     run_primary = config["run_primary"] == "true" || config["run_primary"] == "1" || config["run_primary"] == "yes";
     if (run_primary)
         init_primary();

osd.h

@@ -35,6 +35,10 @@
 #define OSD_PEERING_PEERS 1
 #define OSD_PEERING_PGS 2
 
+#define IMMEDIATE_NONE 0
+#define IMMEDIATE_SMALL 1
+#define IMMEDIATE_ALL 2
+
 //#define OSD_STUB
 
 struct osd_op_buf_list_t
@@ -173,6 +177,7 @@ class osd_t
     int client_queue_depth = 128;
     bool allow_test_ops = true;
     int receive_buffer_size = 9000;
+    int immediate_commit = IMMEDIATE_NONE;
 
     // peer OSDs
 

osd_primary.cpp

@@ -326,6 +326,8 @@ void osd_t::continue_primary_write(osd_op_t *cur_op)
     else if (op_data->st == 3) goto resume_3;
     else if (op_data->st == 4) goto resume_4;
     else if (op_data->st == 5) goto resume_5;
+    else if (op_data->st == 6) goto resume_6;
+    else if (op_data->st == 7) goto resume_7;
     assert(op_data->st == 0);
     // Check if actions are pending for this object
     {
@@ -373,24 +375,67 @@ resume_4:
     op_data->st = 4;
     return;
 resume_5:
-    // Remember version as unstable
+    // FIXME: Check for immediate_commit == IMMEDIATE_SMALL
-    osd_num_t *osd_set = pg.cur_set.data();
+    if (immediate_commit == IMMEDIATE_ALL)
-    for (int role = 0; role < pg.pg_size; role++)
     {
-        if (osd_set[role] != 0)
+        op_data->unstable_write_osds = new std::vector<unstable_osd_num_t>();
+        op_data->unstable_writes = new obj_ver_id[pg.pg_cursize];
         {
-            this->unstable_writes[(osd_object_id_t){
+            int last_start = 0;
-                .osd_num = osd_set[role],
+            osd_num_t *osd_set = pg.cur_set.data();
-                .oid = {
+            for (int role = 0; role < pg.pg_size; role++)
-                    .inode = op_data->oid.inode,
+            {
-                    .stripe = op_data->oid.stripe | role,
+                if (osd_set[role] != 0)
-                },
+                {
-            }] = op_data->fact_ver;
+                    op_data->unstable_writes[last_start] = (obj_ver_id){
+                        .oid = {
+                            .inode = op_data->oid.inode,
+                            .stripe = op_data->oid.stripe | role,
+                        },
+                        .version = op_data->fact_ver,
+                    };
+                    op_data->unstable_write_osds->push_back((unstable_osd_num_t){
+                        .osd_num = osd_set[role],
+                        .start = last_start,
+                        .len = 1,
+                    });
+                    last_start++;
+                }
+            }
         }
+        // Stabilize version sets
+        submit_primary_stab_subops(cur_op);
+resume_6:
+        op_data->st = 6;
+        return;
+resume_7:
+        // FIXME: Free them correctly (via a destructor or so)
+        delete op_data->unstable_write_osds;
+        delete[] op_data->unstable_writes;
+        op_data->unstable_writes = NULL;
+        op_data->unstable_write_osds = NULL;
+    }
+    else
+    {
+        // Remember version as unstable
+        osd_num_t *osd_set = pg.cur_set.data();
+        for (int role = 0; role < pg.pg_size; role++)
+        {
+            if (osd_set[role] != 0)
+            {
+                this->unstable_writes[(osd_object_id_t){
+                    .osd_num = osd_set[role],
+                    .oid = {
+                        .inode = op_data->oid.inode,
+                        .stripe = op_data->oid.stripe | role,
+                    },
+                }] = op_data->fact_ver;
+            }
+        }
+        // Remember PG as dirty to drop the connection when PG goes offline
+        // (this is required because of the "lazy sync")
+        this->clients[cur_op->peer_fd].dirty_pgs.insert(op_data->pg_num);
     }
-    // Remember PG as dirty to drop the connection when PG goes offline
-    // (this is required because of the "lazy sync")
-    this->clients[cur_op->peer_fd].dirty_pgs.insert(op_data->pg_num);
     // Remove version override
     pg.ver_override.erase(op_data->oid);
     finish_primary_op(cur_op, cur_op->req.rw.len);
@@ -483,11 +528,14 @@ resume_2:
         }
     }
     unstable_writes.clear();
-    // SYNC
+    if (immediate_commit != IMMEDIATE_ALL)
-    submit_primary_sync_subops(cur_op);
+    {
+        // SYNC
+        submit_primary_sync_subops(cur_op);
 resume_3:
-    cur_op->op_data->st = 3;
+        cur_op->op_data->st = 3;
-    return;
+        return;
+    }
 resume_4:
     // Stabilize version sets
     submit_primary_stab_subops(cur_op);